ePMP 1000 ePTP drop session and high CPU load

Hi We have a PTP with two ePMP 1000 5 GHz, in ePTP master/slave mode.

It's a very short link, about 200 meters, with RSSI -53 dBm, tx power -4 dBm, SNR 42 dBm and MCS D/U stable at 15. We see that the session, randomly, is lost, even when the throughput is low (30-40 Mbps).

No DFS.

To reduce downtime to 5-6 seconds, there is only one frequency selected in the frequency list.

I don't know if it's correlated, but when the traffic load is at maximum (~95 Mbps aggregated), load average is unusually high, for a bridge installation: 95 %

- edit -

Sorry, forgot to say firmware: 3.0, on both devices. Same problem with 2.6.1.

Curious... we may be having similar issues. What happens when you convert the link to TDD-PTP?

We've had something possibly similar happen with some ePTP links, and we did a test and converted one to TDD-PTP and the problem went away.

We are having same issues with a epmp PTP in eptp mode, about 1,5 miles and drop session at high throughput(80-90Mbps). v3.0


@Eric Ozrelic wrote:

Curious... we may be having similar issues. What happens when you convert the link to TDD-PTP?

We've had something possibly similar happen with some ePTP links, and we did a test and converted one to TDD-PTP and the problem went away.


Later this day I'll give a try with TDD,

Thanks.

Hi,

my last two links had problems with eptp mod , I could not test much because they where very active swiching to TDD solved it also!

It seems switching to TDD PTP solved the drop session issue also for me.

As drawback the higher latency, but this was expected.

CPU load seeems to be slightly higher.

Hi, I got this problem with FW version 3.0.1. I want use ePTP for better latency. Will be in the next version the problemĀ solved? Is problem solvet in beta FW?

Thanks

Mark